Controllerless Networks

Hello everybody,

 

I have an issue with my virtual cluster. It has 103 APs, some of them are IAP-115s and some them are IAP-225s. Now I am trying to add one IAP-225 more and it drops another one from the cluster. This issue repeats itself only with two IAP-225s. I have tried to change these IAPs to new ones.

 

So when I add the new one in the cluster, it drops the second one and gives me following console outputs:

 

Output from IAP-225  which drops from the cluster:

 

[  676.022916] i am a potential master now
[  676.068814] (08:08:10) !!! Slave ---> Pot-Master
[  685.022944] i am a master now
[  685.058427] (08:08:19) !!! Pot-Master ---> Master
[  685.114730] Del bridge entry for magic vlan GW, 18:64:72:c2:57:a6
[  685.187700] asap_send_elected_master: sent successfully
[  709.455449] OOPS. someone else thinks he is the master too, beacon version 4
[  709.539876] I am MASTER. recv-ed a master normal-beacon.
[  709.603475] (08:08:43) master provision, 0 vs. 0
[  709.658724] (08:08:43) !!! Election result, -5
[  710.650013] OOPS. someone else thinks he is the master too, beacon version 4
[  710.734449] I am MASTER. recv-ed a master normal-beacon.
[  710.798047] (08:08:44) master provision, 0 vs. 0
[  710.853291] (08:08:44) Beacon ap-ip, 172.17.7.133 vs. 172.17.7.84
[  710.926250] (08:08:44) Beacon 3G capability, 0 vs. 0
[  710.985653] (08:08:45) Beacon ap-class, 25 vs. 26
[  711.041938] (08:08:45) Beacon mac, 18:64:72:c7:99:e2 vs. 18:64:72:c2:57:a6
[  711.124261] (08:08:45) !!! Election result, 1

 

Output from IAP-225 that I add in the virtual cluster:

 

[   79.752399] asap_send_elected_master: sent successfully
[   79.840885] process `snmpd_sap' is using obsolete setsockopt SO_BSDCOMPAT
[   80.500802] wait for stm to initialize over
[   80.550891] asap_send_elected_master: sent successfully
[  100.776642] ADDRCONF(NETDEV_UP): eth1: link is not ready
[  100.840318] ethernet_device_event: dev eth1 is up
[  105.707463] asap_send_elected_master: sent successfully

 

Also sometimes when I try to add this new IAP-225, the cluster changes its virtual master.

 

I would be more than glad if someone could help me.

 

Best Regards,

 

Eemeli

Make it's connected on port e0 and that it has the same version code .

Try to purge it just in case it was configured to be preferred master

Hi,

 

I have tried all these things allready and none of them seems to help. And all IAPs are connected on port eth0.

 

- Eemeli

Your logs show that E1 is connected and rather than E0.  This puts the AP in Heirchical mode and will kick off an election.

 

100.840318] ethernet_device_event: dev eth1 is up
[  105.707463] asap_send_elected_master: sent successfully

Yeah, I don't know why it says so. I am sure it was connected on e0. I also tried to connect IAP-115 (which has only one enet port) and it didn't work either.

 

And even if it was connected on e1 (which it wasn't), how could it drop the other one from the cluster? And the other one always being the same IAP.

Perhaps I should add one detail. This problem goes both ways: When the other IAP has dropped from the cluster I can reboot it and then the second one drops.